1919] Esterly: Reactions of Various Plankton Animals 77 



that the factor of the rhythm seems not to have much importance, 

 though it probably plays some part in the depth migrations. 



It is probably not well to pursue further the discussion of the 

 matters in this section because the experiments are not complete. But 

 omission of these points would be unwise because a most essential 

 consideration is involved, namely: 



The Relation Between Experimental and Field Work 



After what has been said on this topic in other publications from 

 the Scripps Institute (Michael, 1916, Esterly 1917a, 1917c) it is 

 needless to say much here. It seems to me more apparent than ever 

 that a knowledge of what organisms do in nature is indispensable for 

 an intelligent evaluation of experimental work that is undertaken to 

 obtain a basis for the explanation of behavior under natural condi- 

 tions. An observer cannot learn what is to be explained except from 

 knowledge of the organisms of his interest in nature. Experiment 

 without field work will lead to error. For example, if it were said, 

 on the basis of experiment, that Acartia tonsa would be most abun- 

 dant at the surface by day and least abundant by night, it would be 

 a perfectly fair statement so far as experiments with surface animals 

 go. Yet we know that the reverse is true in nature. Again, the experi- 

 mental results with Calanus would justify one in saying that this 

 animal shows practically no diurnal movement. But our field results 

 have led to a different conclusion, for if any plankton animal has the 

 habit of diurnal migration this copepod has. If experiments point 

 in one direction and field observations in another I believe that the 

 latter are more trustworthy. As has been said by Michael (]916, 

 p. xii), "Experiments . . . reveal only what transpires in a laboratory 

 and are necessarily incapable of revealing what occurs in nature." 



Both field and laboratory work are indispensable and the two 

 should go hand in hand if it is desired to know the natural history of 

 a species. It is practically impossible, however, in many cases to inves- 

 tigate behavior under experimental conditions on account of the size 

 of the animals, trouble in maintaining specimens in captivity, and 

 difficulty in even approaching the conditions of the natural habitat. 

 The last consideration is one of more than passing importance. If it 

 were possible to duplicate nature in the laboratory and then secure 

 the element of experimental control there would be no doubt that 

 behavior in the laboratory would reveal that in nature, in so far as 

 individual behavior can be taken to represent that of a species. 



